Fully Digitalized Optoelectronic Angular Position Sensor and Its Application in Real-Time Gait Recognition

• Published in: JOURNAL OF NANOELECTRONICS AND OPTOELECTRONICS
• Main Authors: Wang, Xijuan
• Format: Article
• Language: English
• Published: AMER SCIENTIFIC PUBLISHERS 2023
• Subjects: Engineering; Science & Technology - Other Topics; Physics
• Online Access: https://www-webofscience-com.uitm.idm.oclc.org/wos/woscc/full-record/WOS:001167716000001

The optoelectronic angular position sensor, as a component for integrated optical-electromechanical angle measurement, operates by converting angle position information into optical signals through a photonic code disc. Subsequently, these optical signals are transformed into electrical signals through photoelectric conversion to measure axial rotation information. This study proposes a fully digitalized optoelectronic angular position sensor, wherein Moore stripes are digitally subdivided, and high-order inner ring angle division employs Gray code encoding. A photoelectric diode array serves as the sensing element, corresponding one-to-one with the encoding channels on the code disc. The output of the photoelectric diodes is binarized through comparator processing, thus converting it into encoded electrical signals. Hardware implementation utilizes PIN photosensitive diodes as sensing elements, designs a laser driver circuit, and employs the 74HCT165 chip for serial-to-parallel conversion. The FPGA program is debugged using JTAG, and the program is solidified on an external EPCS1S18 chip through AS downloading. The LDO chips AMS1117-3.3 V and AMS11171.2 V power the FPGA chip. The CH341 chip is used for interface conversion between the FPGA and PC. In experiments, when the fully digitalized optoelectronic position sensor is installed, the voltage output of approximately 2.2 V meets the sensing response requirements after optical path attenuation. After data transmission, the results are correctly displayed on the host computer. The designed sensor is applied to gait recognition through tests involving single-person walking, two-person walking, and mixed walking. The results show an accuracy rate exceeding 95%, indicating its suitability for gait recognition in footprints.